On: AWS multiple services outage in us-east-1

[binsquare]

The internal disruption reviews are going to be fun :)

[Msurrow → binsquare]

The fun is really gonna start if the root cause of this somehow implicates an AI as a primary cause.

[karel-3d → Msurrow]

I haven't seen the "90% of our code is AI" nonsense from Amazon.

[some_furry → karel-3d]

https://archive.is/b6aUD

[gdulli → karel-3d]

Their business doesn't depend on selling AI so they have the luxury of not needing to be in on that grift.

[aurareturn → Msurrow]

It's never an AI's fault since it's up to a human to implement the AI and put in a process that prevents this stuff from happening.

So blame humans even if an AI wrote some bad code.

[Msurrow → aurareturn]

I agree but then again it’s always a humans fault in the end. So probably a root cause will have a bit more neuance. I was more thinking of the possible headlines and how that would potentially affect the public AI debate. Since this event is big enough to actually get the attention of eg risk management at not-insignificant orgs.

[xenocratus → Msurrow]

> but then again it’s always a humans fault in the end

Disagree, a human might be the cause/trigger, but the fault is pretty much always systemic. A whole lot of things has to happen for that last person to cause the problem.

[Msurrow → xenocratus]

Also agree. “What” build the system though? (Humans)

Edit: and more important who governed the system, ie made decisions about maintainance, staffing, training, processes and so on

[portaouflop → Msurrow]

It’s gonna be DNS

[WelcomeShorty → portaouflop]

Your remark made me laugh, but..:

"Oct 20 3:35 AM PDT The underlying DNS issue has been fully mitigated, and most AWS Service operations are succeeding normally now. Some requests may be throttled while we work toward full resolution."

https://health.aws.amazon.com/health/status

[erpellan → portaouflop]

It’s always DNS! Except when it’s the firewall.

[blibble → Msurrow]

if that's the case it'll be buried